The field of image processing has produced a wide variety of tools for manipulating 3D medical images (volumes) and for extracting meaningful information. Typically low-level image processing tools feed high-level, application-specific analyses, such as segmentation, registration and shape analysis. However, the analysis of brain images, places unique demands on the low-level image processing that underlies these algorithms. New image modalities, application specific image features, and large amounts of 3D data place new demands on the image processing that is necessary for this project. This proposal addresses these specialized demands by focusing on fast methods for extracting feature information from DTI data.